Neural Network Based Automatic Fingerprints Classification Algorithm

It Presented one neural network based automatic fingerprints classification algorithm. Fingerprint features extracted through resolving directed graph are input into neural network to be classified. Weight coefficients of network connection are studied and optimized based on the genetic algorithm. Test results show that its performance is excellent and the total accurate classification rate reaches up to 93.12%. Research in the paper is significant to fingerprint identifications applied in large-scale information security authentication systems such as bank saving network system and network exchange system.